The Future of Organoid Intelligence

Todd Poley (00:00):
Hi, and welcome to Bold New Approaches, a limited series of the Vital Science Podcast from Charles River. I'm your host, Todd Poley, and in this series we will bring you innovators and leaders that are shaping the evolving world of drug discovery and development. Our aim is to spotlight this connecting bridge of reputable battle-tested science with the rapid advancements in emerging technologies. We'll take a close look in alternative methods that responsibly reduce animal impact and equally create efficiencies and new insights for drug developers. Our first guest of Bold New Approaches is exactly that, an innovator, an accomplished and globally recognized toxicologist and a disruptor to the traditional methodologies within in vivo testing and looking towards the promising potential of organoid cultures combined with artificial intelligence. This individual is the former head of European Commission's Centre for Validation of Alternative Methods, has over 560 scientific publications and has received many awards and accolades from his peers. Please welcome Dr. Thomas Hartung. Welcome Thomas. Thanks for joining us on Bold New Approaches.

Dr. Thomas Hartung (01:15):
Thanks for having me.

Todd Poley (01:17):
So maybe to start off, Thomas, tell us a little bit about your current roles and what you've been focusing in on.

Dr. Thomas Hartung (01:26):
Yeah, I mean I'm an academic and you mentioned already my short pause of seven years where I was moving to the European Commission and tried to support policy by science focusing on what has been for 35 years my quest, which is replacing animal testing by more modern technologies because you have to understand, I just turned 60. Most of the methods which we use, especially in safety science and toxicology have been introduced when I was not yet born or in kindergarten. But after a while I thought, okay, academia freedom, the possibility to speak without 10 permissions to talk to journalist are nicer. And so I've really embarked in academia again. I'm holding now five professorships at three universities. So I'm really broad and you can also take from this, it's not necessarily deep. I'm really spanning from pharmacology, toxicology, engineering, infectious diseases and because I like to combine things and this is what I'm doing.

Todd Poley (02:40):
Yeah, that's really, I think you challenge the status quo where it's almost the revolution in toxicology where this 50 plus year old animal testing methodologies to what you have been focusing and your love for organoids. Would you explain to our audience what organoids are and maybe organoid intelligence?

Dr. Thomas Hartung (03:02):
Yeah, absolutely. And you're right, I was most proud when I was introduced to talk to the OECD, which is the repository for all of the important test methods. They introduced me from the OECD, Thomas is steering a revolution in toxicology. This was what I really wanted to hear. It was like a knighthood.

Todd Poley (03:23):
I'm sure.

Dr. Thomas Hartung (03:25):
It is about doing things different. And when the idea of replacing animal testing, which was the fundamental tool and almost the only tool until the 70s, when this idea was proposed, it was a vision, but nothing but a vision. There was nothing really to replace animal testing. And it is only in more recent time that things come into reach, which are really disruptive, which are really offering something which is at least as good if not better than the animal test because most of the cell culture, which was always put forward, these were tumors of humans which had been taken out. And then a single cell of these was transformed into a cell line and these cell lines were growing like wheat, very different to what is happening in the body, but they also have not a lot in common with cells. And it was really only with the advent of stem cells that we could produce high quality human tissues.

I mean, animal tissues is not the same problem. You get access to this, but imagine I want to work with human cells and I have to work with them because more than half of the new drugs are actually human proteins or antibodies against human proteins and we simply cannot use animal models, but ask a human to give something perhaps some blood, perhaps a little bit of skin, but that's it. Yeah. I'm working on the brain, to pick someone's brain. I'm not very successful in asking. So you can imagine this was really only when stem cells became available that we could start engineering these tissues. And this opened up for what we now call an organoid, which means it's a tiny replication of an organ, but it's not just a single cell. It is something that shows some architecture and some functionality of the organ, and this is what we call an organoid.

Todd Poley (05:39):
When you speak of organoid intelligence, what does that mean? Does that mean that these little animals are sentient beings? Are they conscious? I mean, what exactly does that mean?

Dr. Thomas Hartung (05:51):
Oh, making a big leap coming from organoids to organoid intelligence, but what I just said is I'm interested in brain. And so for us it's just brain organoids and we were really at the forefront. You have to imagine the cells allowing all of this have become available in 2006. In 2013, the first time somebody described how to use these cells and produce an organoid, and we were just four months behind them, behind this nature publication. So we are really working at the very same time on these things and our first contribution was to mass produce brain organoids. So highly standardized actually little balls of cells, but showing a lot of the cell types, showing that these cells are communicating as they do in the brain. They form circuits, they're talking with electrical signals to each other. So this was the first step, and we have been using this now for a broad number of diseases from infections to tumors to toxicity.

We're trying to find out what is promoting autism and other diseases. But I said earlier, the big advance is you have functions. So what is the function of the brain? It is not just to survive and live and communicate senseless with each other among these cells, it is about computing, getting inputs and producing outputs. And so I often said when I saw these cells actively communicating, I said, "They're thinking." And then some people say, "Oh, are they conscious?" And I said, "No, no, don't worry. They have nothing to think about. They have no input and no output."

Todd Poley (07:44):
Very interesting.

Dr. Thomas Hartung (07:45):
The outcome, organoid intelligence. Organoid intelligence was born out of this, very easy way of putting it. I asked myself at some point what is happening if we give them input and output? And this created something we now call organoid intelligence where we really try to make such a cell culture a little bit like a tiny computer, not a computer,-

Todd Poley (08:12):
So a tiny computer powered by cells. That's amazing.

Dr. Thomas Hartung (08:14):
Yeah.

Todd Poley (08:14):
Amazing.

Dr. Thomas Hartung (08:15):
So the idea was very simple. What happens if these cells are not bored to death, but we use some electrodes to on the one hand record what they're doing, their communication and on the other hand we are feeding them with some information. And this was giving rise to organoid intelligence. There was an opportunity for big grant application and I assembled a team here in Hopkins with some of the superstars. Everything I needed here, the bioengineers, the cell biologists. And then very importantly, the artificial intelligence is giving us at this moment miracle tools.

Todd Poley (09:00):
That's amazing. Yeah.

Dr. Thomas Hartung (09:01):
AI is being used to talk to animals now. I was trying to talk to brain organoids and this is how organoid intelligence was formed, as the fusion of organoids and AI as a synthetic biological intelligence.

Todd Poley (09:18):
So let's talk more about that, the transition, the opportunity to focus from an in vivo to an in vitro origination. So in what ways can culturing organ specific tissue from stem cells change the way that diseases are studied and treated compared to what they have been?

Dr. Thomas Hartung (09:40):
I mean, the first thing is that we are not two-dimensional. Yeah. So we are not flat animals. We are three-dimensional and a three-dimensional cell culture is simply much more adequate. This really starts with the fact that we have about 1000 times higher cell density in a tissue than we have in a cell culture. If you imagine cell cultures, regular cell cultures, they look like pan fried egg, sunny side up. The egg yolk is really the nucleus. Yeah. So these cells are flat, they have hardly any contact to each other. 50% what they see is plastic and 50% is cell culture media, completely unphysiological situation. And then take cells like human neurons, which on average are linked to 8,000 other neurons. So each neuron has 8,000 connections. So how should I do if they're laying flat and spread out? So they have much, much less connections, is the most simple thing.

And so by making this a three-dimensional culture, we already enable the connections which are necessary for learning for example. The learning process also is not a thing which is only done between the neurons. We have in the brain beside neurons, what was often belittled as helper cells, the astrocytes and oligodendrocytes or the microglia mainly. But these cells play very fundamental roles. They make the signals go faster, they support the neurons, but most importantly in our context for cognition and intelligence and memory, they are the ones which are pruning the connections which are not needed and strengthening those which are having the information. So it is only with the availability of such systems which suddenly have all of the players that we can start thinking about establishing any cognitive functions.

Todd Poley (11:52):
So I want to ask you though, this is really, I mean the potential is enormous and thinking about this evolution from traditional animal testing to this new potential, what drove you to focus on organoid intelligence as a toxicologist?

Dr. Thomas Hartung (12:12):
I mean, we were very much motivated by autism, the increase of autism. In the US, this is very well documented because the Center for Disease Control does an excellent job in doing so. You have to imagine autism was diagnosed in one in 10,000 children in the 70s. Now the latest data from last year were one in 36 children is diagnosed with autism. And while in the beginning there was a lot of more awareness and setting of diagnosis, we cannot say so for the last one or two decades. It's really using the very same criteria and high alert for identifying this. We see a continuous increase. There's a genetic component to autism. We know this, but genetic cannot change that fast. We need human models. I mean there's no autistic rat, there's no rat in the corner communicating with the other rats. Yeah.

This is a problem. You need a model. And many things, especially if they concerned the brain are quite unique in humans. And then added to this that the new entities we want to test are, as I said earlier, human antibodies or proteins from humans. They do not work in the animal models. So to have something which is possibly reflecting what is going wrong in this process is an example. And this was our motivation, our starting point. But if you look, what is affected in these children? It is cognition. It is not that cells die. They don't have smaller brains because a lot of certain population of cells is gone. It is about the connections they form between these brain cells. Exactly what is making up the long-term memory. This is not a memory defect to be clear. It's in this case it's behavioral effect, but if we want to have readouts which are relevant, we need to somehow emulate the functions of these organs.

Todd Poley (14:31):
So progress is being made. I think we've heard programs like organ on a chip technology that is aiming to facilitate earlier in vitro testing of a therapeutic effect. And we talked a little bit about mimicking human physiology, which I think is a big part of this, right? So what other benefits can we see from the drug development industry? And maybe the other part of that question would be how can this disrupt the current drug discovery and development pipelines?

Dr. Thomas Hartung (15:05):
Yeah, you have to imagine our organs have blood circulation. They're supplied with oxygen and nutrients in the body continuously. And through this perfusion of the organ, we have homeostatic conditions, which means nothing is changing. The cells are permanently supplied with all of these and everything is stable. Our cell culture is completely different. We add far too much nutrients, highly diabetic sugar concentrations to the cells. And then over the next two or three days, they're consuming this based is increasing oxygen in the beginning a lot, getting less and less. And then in a second we are changing these media and they have to adapt. And this is really something which is perturbing the quality of our cell culture dramatically. So the out is to create perfusion. There's a lot of technological solutions, hollow fibers and hydrogels and whatever name it, this is not important. But we have the technologies to really allow stable conditions for these cells and continuously provide them with what they need.

And this has enormous impacts. It allows us, for example, to grow larger brain organoids. A brain organoid cannot be made larger than about half a millimeter. If we go beyond this, the oxygen and the nutrients don't get fast enough into the center of this and they start rotting. And this is not what you want for testing because then the organoid is struggling with its decay and is impacted by this. You want a healthy brain organoid, especially if you want more complex functionalities. So on the other hand, we have 180 billion cells in a human brain. The maximum we can do is 50 to 100,000 in such a brain organoid without a perfusion. So you see the scale. I mean, brain is something where size matters. Our brain organoids, which we are producing are standardized in a standardized way, have as many neurons as a housefly. So you really want something bigger. So perfusion is already prerequisite for building something large enough worth training.

Todd Poley (17:42):
So you spoke to autism.

Dr. Thomas Hartung (17:45):
Yeah.

Todd Poley (17:46):
What other disease areas can this approach impact on the maybe more near future?

Dr. Thomas Hartung (17:53):
In the beginning I was quite skeptical about the big diseases of the brain we are looking at like for example, Alzheimer's disease, which is in desperate need where the lack of adequate animal models is one of the reasons we're making so little progress. We're seeing that these artificial conditions we are creating to produce something Alzheimer-like in animals does not really work well and there's not much on the market really. But Alzheimer's something which happens after 50, it's very rare that you have early forms of dementia. Can you really, with a system which is made up from embryonic cells developing for a few months, can they really show something? But the interesting finding was that as soon as you take these cells from an Alzheimer patient and reprogram them to become stem cells, they show some of the features of Alzheimer. They show the histopathology, the cell structure, the tissue structure, which we typically see in Alzheimer patients.

So if we now do the next step, this has not yet been done, but this is one of the things we are at the moment approaching is our Alzheimer center here in Hopkins. We want to see does a brain organoid, which shows these defects cellularly, does it also show the defects of an Alzheimer patient? It is not necessarily a memory problem here. It is really more a problem of plasticity. Do the cells as easily reconnect and change functionalities? This is what we are trying to address. We call this Dory after this fish from Finding Nemo had no long-term memory we recall. It's the Dementia Originate Intelligence Research Initiative. So we are building up a systematic program at the moment to bring these technologies together, the organoid intelligence as the readout and the creation of cellular aspects of dementia in the Alzheimer research. But you can imagine just the same. Yeah. We do a lot of work on virus infections of these.

For example, we were the first on May of 2020 already to show that human neurons can be infected with COVID-19 virus, SARS-CoV two and is this part of long COVID? We see that certain low-level brain infection is happening in patients. It could very well be that these things are impacting on cognition or ability to concentrate, to learn the dizziness these people feel in long COVID. This is just one example of what you can do.

Todd Poley (20:50):
Yeah.

Dr. Thomas Hartung (20:50):
On the safety side, there's tons of opportunities. We know that like autism, Alzheimer's is increasing dramatically, more than the aging of the population would suggest. So it is not just that we see so much more Alzheimer's disease now because we have more old people. No, there's something else, but what is the else? You need a test system. It could again be the same suspects. If something is perturbing the development of the brain, it could as well also accelerate the degeneration of the brain. If we have a test system, let's say of an Alzheimer patient, then we know the genetic is there to develop this, then we could look, are pesticides accelerating this, is cannabis and accelerating this, is flame retardant accelerating this? Just to mention a few of the possible culprit's nobody has been able to study before.

Todd Poley (21:56):
So for the drug discovery and development programs out there, how can this be expedited as more of a accessible standard research option?

Dr. Thomas Hartung (22:07):
You have to see for the drug development process. It is all about time to market and to narrow it down to the most promising substances. To develop a new drug is a more than 2.4 billion process. I say more because it was 10.4 billion in the last assessment I found 10 years ago. So it is incredibly expensive and on average last 12 years, which means there's only about eight years of patent runtime left to make the money. Or you can say it another way. One day delay to getting to the market is worth about $2 million just to break even with the development costs. Okay. What does it mean? Every company in the world is looking for the most efficient and fast strategy to get to something. Animal models are anything but fast. It takes quite a while to set them up. These cell culture type of experiments are typically very short.

They come to results within weeks, not within months or years. For this reason, they're already promising to accelerate. And they are also very often robotizable, you can automate these. You can test relatively many substances. There's now already organoids on 384 wheel plates, which means on such a small cell culture plate you can test 384 conditions. So there's a lot of opportunity of using them between the very early phase of what we call high-throughput screening. There are really millions of substances often not being tested before you go then into the highly complex disease models as a very meaningful filter. And that's where I see these models mainly. So this is for the development side.

And now take that about 25% of all clinical trials are for brain diseases. So there's essentially no, this is a really important part of it. And then even those which are not meant for the brain, you want to de-risk them for the brain because what are the side effects which very often lead to not continue, not being able to continue a program. These are things like, simple things like dizziness or nausea, which is being caused. Many things happen in the brain. Pediatric trial, one of the big concerns is that brain development of the child could be impacted if you're treating with a certain thing. If we can de-risk and say no, has no effect on brain functions, this would be a big step forward.

Todd Poley (25:02):
It's amazing to see the potential when you fuse this machine learning AI boom that we've been seeing over the past several years with traditional models. What one piece of guidance would you want to share with the biopharmaceutical industry? Maybe a brief statement that you'd like to leave the audience with.

Dr. Thomas Hartung (25:29):
I mean, we have not really talked about the AI side of things, which you just mentioned. I've been very much invested in this. I'm among others the field chief editor of Frontiers in AI, and we have published nine articles in this field. And we have been using AI and only with AI, it is possible really to have the data analytics of what such a brain system is producing. You have to imagine we are putting our brain organoids at the moment on 28,000 electrode arrays, yeah, which are continuously producing signals. We're really talking here about terabyte of data which are being produced and it can only make sense, fast enough with the AI system. And so organoid intelligence would not be possibly without this AI part. And we were in the fortunate situation that we have a certain opinion leadership on both sides and could engage colleagues who have more understanding than we because of this spark, which comes from the combination.

But you asked me to give a final statement. What I think is really important to understand, we do have disruptive technologies which come on the one hand out of the bioengineering field and advanced cell culture. And on the other hand, from the information technologies. These are opening up for a completely different paradigm to optimize a drug development process, which is at the moment almost broken. We see that it's extremely difficult to refill the pipeline with ultimately successful drugs. And these are two tools which are fundamentally different to what we have been using in the past. And we have seen that in 2022 already 18 AI first drugs came into clinical trials last year. The first AI designed drug came into a phase two trial, and now imagine if you at the same time enhance this with human cell systems, which can give you in a predictive way, change the odds of success. That's where I really see promise.

Todd Poley (27:49):
Yeah, it's almost parallel benefits at the same time, right? It's therapies that are brought to the patient more efficiently, better translatable aspects along with it and reducing animal impact along the way is really a win-win for everyone. So great wisdom. Thank you for sharing. Dr. Thomas Hartung, thank you so much for being part of our first episode of Bold New Approaches, which is part of our limited series for Vital Science. Thanks for coming on and sharing that.

Dr. Thomas Hartung (28:24):
Thanks for having me.